Political Science Educator: volume 28, issue 2
Reflections
By Debra Leiter (leiterd@umkc.edu) and Danielle Joesten Martin (danielle.martin@csus.edu)
Everyone loves a nail-biter, a close competition where the winner is left in doubt until the very last minute. When the stakes are high, people care about winning and everyone pays attention. We are, of course, not referring to the 2024 US presidential election or the 2024 Super Bowl (with a breathtakingly close game narrowly won by the Kansas City Chiefs in overtime over the San Francisco 49ers), but the contest between the University of Missouri-Kansas City Political Science students versus those of California State University-Sacramento for who could most accurately predict the 2024 Super Bowl winner. Our in-class Super Bowl forecasting assignment is a short-term lesson that not only allows students to competitively forecast a high-stakes public event, but also helps students understand political science methods in alternative contexts.
Forecasting is an important element of political science–running the gamut from elections to conflict–and an effective teaching tool for linking theoretical concepts with real-world phenomena. However, teaching forecasting is challenging in introductory classes, as many forecasts are methodologically intensive and require a knowledge of key political science theories. While election forecasting (Berg and Chambers 2019; Leiter 2023; McGee and Hall2024) and event forecasting (Robertson 2024) assignments exist, there are relatively few short forecasting assignments that do not take too much class time and can easily fit in non-methods classes. We think that Super Bowl forecasting is a useful solution.
The overlap between election and athletic forecasting is greater than it appears. Calculating the expected spread of a game uses similar logic to ensemble election forecasting (Graefe et al. 2014). And just like betting on sports outcomes, in some countries interested observers can place bets on election outcomes–a method used extensively (e.g., Graefe 2024; Perez 2023; Wall, Sudulich, and Cunningham 2012) and with high levels of press coverage (Garrett 2024). Political scientists have written extensively on the connections between athletics and politics (Bailey and Trantham 2021; Scheiner 2023; Wallsten, Nteta, and McCarthy 2022), and many prominent media forecasting figures like Nate Silver began as sports forecasters (Butterworth 2014).
Super Bowl forecasting, like other types of forecasting assignments (McGee and Hall 2024), is also an excellent tool for engaging students, especially general education students, who often start with less interest and connection to politics than political science majors (Kalaf-Hughes 2021). The Super Bowl is useful for engaging with students for three reasons First, many students are already interested, regardless of their major, especially if a local team is participating (as was the case for us[1]). Even without that connection, more than 123 million people watched the 2024 Super Bowl, compared to the 67 million who watched the Harris-Trump debate (Akabas 2024). Indeed, even those not particularly interested in football tune in to the Super Bowl for the halftime show, viral commercials, and food and socializing at Super Bowl parties. The second advantage is having students see the connections between course material and the wider world. Making the comparison to sports makes political concepts more relatable, as students have had frequent opportunities to watch or play sports and games, and in turn, helps them see the application of these theories to contexts outside those discussed in class, and even potentially see their own role in applying these concepts (Bar-On 2017; Dreyer 2013). The final advantage is scheduling. One challenge of teaching election forecasting is that there needs to be an event that students can reasonably forecast; in the wrong semester, there may not be an election that your students have enough information about to forecast successfully. The Super Bowl’s date is known well in advance, however. Compared to many other athletic championships, the Super Bowl is easier to model because it is decided in a single event, rather than in a series, meaning students can make just one forecast and immediately see results.
Assignment Instructions:
- Explain the logic of forecasting[2] to students.[3] Discuss how forecasts can come from many different sources, and how models are assembled by weighting various factors to create a prediction.
- Provide students various sources of information about the Super Bowl and allow them time to look up additional information. For example, students can be given a spreadsheet of all previous Super Bowl games points[4] (Super Bowl Winners by Year – ESPN n.d.) or access to team statistics from the season.
- Have students complete a short survey making a prediction about how many points each team will get, along with their confidence in their predictions, their football knowledge, and their team affiliation (the survey we used is available here[5]). If you are running the forecast as a competition between universities or classes, you can have all students complete the same survey and report their class and/or university affiliation–this allows for easier comparisons between forecasts.
- Calculate the class forecast for the Super Bowl once you download and enter student forecasts for each team and their level of confidence. As the purpose of this assignment is to get students to understand the underlying logic of forecasting and weighting, we calculated two simple forecasts: a raw average and an average weighted by self-reported confidence[6].[7]
Following the calculations, present the models to the students so that they can see exactly how you created the forecast and what the predictions are. We were able to show students how house effects worked, given the stark differences in the direction of the predictions of the two university classes, along with the ways that weighting changed the overall quality of the prediction. You can return to these results and differences following the Super Bowl to show students how the results reflected reality.
Crafting Competition:
While there are many ways of engaging students in the material, making an assignment a competitive game has been shown to be an effective strategy (Kollars and Rosen 2017). To increase the stakes, our classes (one in Kansas City, home of the Chiefs, and one in Sacramento, rooting for the San Francisco 49ers) challenged each other to see who could most accurately forecast the game. We posted our predictions on social media[8], and the winning class received a prize from the losing class (“My class won the 2024 Super Bowl forecast, Political Science Edition” stickers, which winning students proudly wore). We also created within-class competition by giving a prize (for example, extra credit in the class) to the student who had the most accurate forecast, so students were able to have two layers of competition. By creating a competition, the assignment replicates the general experience of forecasters in real world environments, where correct forecasts can have dramatic economic, policy, and political consequences.
By giving students the opportunity to forecast the Super Bowl, and turning it into a public competition between classes, students are moved from the sidelines to calling their own play, allowing them to test their understanding of political methods immediately and in new contexts. In a short amount of class time and with limited calls on the faculty time, we think this wild card assignment will have students calling for an instant replay.
Endnotes
[1] We were able to arrange our competition for this because we had attended graduate school together, so we had a pre-existing connection. However, we think these kinds of assignments are also a great way to build pedagogical partnerships with other universities and faculty members and encourage faculty to reach out.
[2] https://educate.apsanet.org/resource/09-07-2022/election-forecasting-assignment-and-teaching-resources
[3] Leiter (2023) has created teaching material on election forecasting, available on the linked page at APSA Educate.
[4] https://docs.google.com/spreadsheets/d/1x2BvoAsHVOY0PbBQmOMxTBNjC_89rlf9/edit?usp=share_link&ouid=113756150849823742282&rtpof=true&sd=true
[5] https://docs.google.com/document/d/1a2hztuW1IEBJXMK9zQEz_qIVrRAqOi71OmkpTIVn4L8/edit?usp=sharing
[6] https://docs.google.com/spreadsheets/d/1uDoi6aHutBu37iuxuQOuECRSAXWOu8Ap/edit?usp=share_link&ouid=113756150849823742282&rtpof=true&sd=true
[7] For the weighted average, we multiplied each student’s score prediction by their confidence level (1-5, where 5 is most confident). So, if Student A predicts a score of 10 for the Chiefs with a confidence of 1, then Student A’s forecast is 10. If Student B predicts a score of 20 with a confidence of 5, Student B’s estimated score is a 100. Then add together the adjusted, weighted scores for the students, and divide by the sum of the confidence scores, essentially giving high confidence students’ estimates five times the weight. Walking students through this logic gives them a simplified overview of survey weighting. This method builds on the logic of the wisdom of the crowds, which has played a robust role in theories of citizen forecasting (Lewis-Beck and Stegmaier 2011; Murr 2011). This system is adjustable to the methodological experience of your students and the theoretical models presented in class.
[8] https://twitter.com/DebraLeiter/status/1756028517014241572
References
Akabas, Lev. 2024. “Trump-Harris Debate Outdraws All NFL Games Besides Super Bowls.” Sportico.com. https://www.sportico.com/business/media/2024/presidential-debate-trump-harris-ratings-nfl-comparison-1234797025/ (November 26, 2024).
Bailey, Kendall L., and Austin Trantham. 2021. “Racial Politics and the Presidency: Analyzing White House Visits by Professional Sports Teams.” Social Science Quarterly 102(2): 897–919. doi:10.1111/ssqu.12944.
Bar-On, Tamir. 2017. Beyond Soccer: International Relations and Politics as Seen through the Beautiful Game. Rowman & Littlefield.
Berg, Lukas, and John Chambers. 2019. “Bet Out the Vote: Prediction Markets as a Tool to Promote Undergraduate Political Engagement.” Journal of Political Science Education 15(1): 2–16. doi:10.1080/15512169.2018.1446342.
Butterworth, Michael L. 2014. “Nate Silver and Campaign 2012: Sport, the Statistical Frame, and the Rhetoric of Electoral Forecasting.” Journal of Communication 64(5): 895–914. doi:10.1111/jcom.12113.
Dreyer, David R. 2013. “Exploring the Concept of Rivalry: From India and Pakistan to the Yankees and Red Sox.” Journal of Political Science Education 9(3): 308–19. doi:10.1080/15512169.2013.796238.
Garrett, Luke. 2024. “Americans Bet $100 Million on Trump v. Harris, but at What Cost?” NPR. https://www.npr.org/2024/10/29/nx-s1-5132616/election-day-betting-trump-harris (January 20, 2025).
Graefe, Andreas. 2024. “The PollyVote Forecast for the 2024 US Presidential Election.” PS: Political Science & Politics: 1–6. doi:10.1017/S104909652400101X.
Graefe, Andreas, J. Scott Armstrong, Randall J. Jones, and Alfred G. Cuzán. 2014. “Combining Forecasts: An Application to Elections.” International Journal of Forecasting 30(1): 43–54. doi:10.1016/j.ijforecast.2013.02.005.
Kollars, Nina, and Amanda M. Rosen. 2017. “Who’s Afraid of the Big Bad Methods? Methodological Games and Role Play.” Journal of Political Science Education 13(3): 333–45. doi:10.1080/15512169.2017.1331137.
Leiter, Debra. 2023. “Teaching Forecasting Without Teaching Methods.” Journal of Political Science Education 19(2): 185–94. doi:10.1080/15512169.2022.2116333.
Lewis-Beck, Michael S., and Mary Stegmaier. 2011. “Citizen Forecasting: Can UK Voters See the Future?” Electoral Studies 30(2): 264–68. doi:10.1016/j.electstud.2010.09.012.
McGee, Zachary A., and Precious D. Hall. 2024. “Using Prediction Markets as a Tool for Classroom and Civic Engagement.” Journal of Political Science Education 0(0): 1–24. doi:10.1080/15512169.2024.2385366.
Murr, Andreas Erwin. 2011. “‘Wisdom of Crowds’? A Decentralised Election Forecasting Model That Uses Citizens’ Local Expectations.” Electoral Studies 30(4): 771–83. doi:10.1016/j.electstud.2011.07.005.
Perez, Vanessa M. 2023. “How the American Public Perceived Electoral Competition in the States during the Pre-Poll Era: A Prediction Market Data Analysis of the 1896 Presidential Election.” State Politics & Policy Quarterly 23(1): 48–67. doi:10.1017/spq.2022.14.
Robertson, Justin. 2024. “Can Chat GPT and Crowdsourced Forecasting Help Students Think About International Relations? A New Class Assignment.” Political Science Educator 28(1). https://educate.apsanet.org/can-chat-gpt-and-crowsourced-forecasting-help-students-think-about-international-relations-a-new-class-assignment (December 3, 2024).
Scheiner, Ethan. 2023. Freedom to Win: A Cold War Story of the Courageous Hockey Team That Fought the Soviets for the Soul of Its People—And Olympic Gold. Simon and Schuster.
“NFL History – Super Bowl Winners.” ESPN.com. https://www.espn.com/nfl/superbowl/history/winners (December 3, 2024).
Wall, Matthew, Maria Laura Sudulich, and Kevin Cunningham. 2012. “What Are the Odds? Using Constituency-Level Betting Markets to Forecast Seat Shares in the 2010 UK General Elections.” Journal of Elections, Public Opinion and Parties 22(1): 3–26. doi:10.1080/17457289.2011.629727.
Wallsten, Kevin, Tatishe M. Nteta, and Lauren A. McCarthy. 2022. “Two Sides of the Same Coin? Race, Racial Resentment, and Public Opinion Toward Financial Compensation of College Athletes.” The Forum 20(1): 63–85. doi:10.1515/for-2022-2049.
—-
Dr. Debra Leiter is an associate professor of political science in the Department of Political Science and Philosophy at the University of Missouri-Kansas City. Her research focuses on comparative political behavior, and her pedagogical research focuses on student research engagement and participation.
Dr. Danielle Joesten Martin is an associate professor of Political Science at California State University, Sacramento. Her research focuses on elections and voting behavior in the United States.
Published since 2005, The Political Science Educator is the newsletter of the Political Science Education Section of the American Political Science Association. As part of APSA’s mission to support political science education across the discipline, APSA Educate has republished The Political Science Educator since 2021. Please visit APSA Educate’s Political Science Educator digital collection here.
Editors: Colin Brown (Northeastern University), Matt Evans (Northwest Arkansas Community College)
Submissions: editor.PSE.newsletter@gmail.com
